Philosophical machines under incomplete knowledge: BOIS, SIMA, BORIS, substrate physiology, and accountable AI

AI governance often begins with outputs, risks, and values, but a deeper operational layer is decisive: the machine that defines objects, evidence, agency, risk, closure, and transition. This article updates the BOIS protocol under current BOIS/BORIS v2.6 by adding SIMA as the analytical layer and BORIS as the implementation layer. BOIS is treated as a transition-oriented philosophical machine for organized consequential activity under incomplete knowledge, not as a universal theory or field-validated product. SIMA reconstructs candidate machines from substrate physiology through opers, operome inventories, and operature correspondence. BORIS builds narrow domain implementations that must be tested by experience. The substrate rule M@S remains central: the same abstract machine, run on different relevant substrates, may preserve a high-level function while acquiring different implementation physiologies.